铜胁迫下玉米叶片光谱奇异性分析及污染评估

发布时间：2018-07-23 21:03

【摘要】：为了区分不同浓度铜离子(Cu~(2+))胁迫引发玉米的叶片光谱间微弱信息差异和铜污染程度,根据玉米叶片实测的SVC高光谱数据和同步获取的玉米叶片中铜含量数据,利用Daubechies小波系中的"Db5"小波,得到玉米叶片光谱在350~2500nm波段范围小波分解的第5层高频组分(d5),并采用盒维法计算d5的分形维数,通过分形维数邻近变化率(α)讨论不同铜胁迫梯度下玉米叶片光谱分形维数的变化趋势,从而可对d5的奇异范围、奇异幅度等光谱奇异性参数进行定量计算和分析.结果表明:d5能精确地探测到玉米不同铜胁迫梯度污染的光谱奇异性弱信息,并实现不同污染程度玉米高光谱信号的分离;d5分形维数随污染程度的增加先减小后缓慢升高,最后达到峰值,其中Cu(100)分形维数值最小;CK(0)与Cu(100)间α为负,其他两个胁迫梯度区间内均为正,且Cu(100)和Cu(300)间α绝对值最小,而Cu(300)和Cu(500)间α绝对值最大;通过建立玉米叶片铜含量估算模型验证了玉米叶片中铜含量与奇异幅度及分形维数之间存在较强的相关关系,各胁迫梯度玉米叶片中Cu含量的差异均达到显著水平(为0.05),其决定系数R2=0.9501.所以,光谱高频组分分形维数和奇异性特征能够有效地诊断并定量分析玉米铜污染程度状况,从而可为作物重金属污染监测提供借鉴意义.
[Abstract]:In order to distinguish the weak difference of spectral information and the degree of copper pollution in maize leaves caused by different concentrations of Cu ~ (2) stress, the SVC hyperspectral data of maize leaves and the data of copper content in maize leaves were obtained simultaneously. Using the "Db5" wavelet in the Daubechies wavelet system, the fifth layer high frequency component (d5) of maize leaf spectrum in the range of 350~2500nm band is obtained, and the fractal dimension of d5 is calculated by box dimension method. The variation trend of spectral fractal dimension of maize leaves under different copper stress gradients was discussed by means of fractal dimension adjacent change rate (伪). The singular range, singular amplitude and other spectral singularity parameters of d5 could be calculated and analyzed quantitatively. The results showed that the spectral singularity of different Cu stress gradients in maize could be detected accurately by the ratio d5, and the fractal dimension of fractionation of hyperspectral signals of maize with different pollution degree was decreased firstly and then increased slowly with the increase of pollution degree. The value of Cu (100) fractal dimension is negative between CK (0) and Cu (100), the other two stress gradients are positive, and the absolute value of 伪 between Cu (100) and Cu (300) is the smallest, and the absolute value between Cu (300) and Cu (500) is the largest. The correlation between copper content in maize leaves and singular amplitude and fractal dimension was verified by establishing a model for estimating copper content in maize leaves. The difference of Cu content in maize leaves was significant (0.05), and the coefficient of determination was R20.9501. Therefore, the fractal dimension and singularity of high-frequency components of the spectrum can be used to diagnose and quantitatively analyze the copper pollution in maize effectively, which can be used as a reference for heavy metal pollution monitoring of crops.
【作者单位】：中国矿业大学(北京)地球科学与测绘工程学院;
【基金】：国家自然科学基金项目(41271436) 中央高校基本科研业务费专项资金(2009QD02)
【分类号】：S513;X826

【相似文献】